A method for source region shrinkage and localization of water quality peaks during river navigation

Abstract

This application discloses a method for source region narrowing and localization of water quality peaks during river navigation, relating to the field of water environment management technology. The method includes the following steps: acquiring continuous multi-parameter water quality data and corresponding location information along the river; identifying continuous segments with abnormally high water quality parameters as water quality peaks using the continuous multi-parameter water quality data and corresponding location information; and extracting characteristic parameters of the water quality peaks, including peak width and the arrival order of multi-parameter peaks. This application constructs multi-dimensional constraints by fusing water flow direction, peak width, and the arrival order of multi-parameter peaks, achieving a step-by-step narrowing of candidate source regions. This transforms the fuzzy location of pollution sources from a large-scale area to a specific spatial segment, improving the targeting and stability of the location. Simultaneously, it directly analyzes the migration process based on the characteristic parameters of the water quality peaks, eliminating the need for complex models, reducing computational dependence, and supporting rapid response and on-site decision-making.

Description

Technical Field

[0001] This application relates to the field of water environment management technology, specifically to a method for source region shrinkage and positioning of water quality peaks during river navigation. Background Technology

[0002] With the increasing demand for water environment governance, the ability to quickly and accurately locate pollution sources has become a key technical issue in emergency response and daily monitoring of river water pollution incidents. In recent years, mobile water quality monitoring technology based on unmanned or manned vessels equipped with multi-parameter water quality sensors has been widely applied. This technology can continuously acquire high-frequency, multi-indicator water quality data (such as ammonia nitrogen, permanganate index, conductivity, and turbidity) along the river and identify abnormal pollution areas through spatial distribution analysis. In actual monitoring, when the monitoring trajectory traverses polluted water bodies, it often forms obvious "water quality peaks" in the data sequence, i.e., continuous sections where water quality parameters significantly increase within a certain spatial range.

[0003] Existing technologies typically rely on the characteristics of such peak clusters, combined with spatial variation trends of pollutant concentrations, simple convection-diffusion models, or empirical attenuation laws, to inversely infer the location of pollution sources, thereby achieving preliminary pollution source tracing and localization.

[0004] However, existing pollution source tracing methods based on mobile water quality peaks still have significant shortcomings:

[0005] First, most methods rely solely on pollutant concentration thresholds or single attenuation models for reverse flow calculations, which are difficult to adapt to the complex and variable hydrodynamic conditions in actual rivers (such as backflow in bends, tributary backflow, and local stagnation), easily leading to an excessively large range of pollution source location with low accuracy.

[0006] Secondly, existing technologies generally neglect the physical significance of water quality peak morphology, especially the degree of pollutant diffusion and emission characteristics reflected by the "peak width," failing to translate it into an effective constraint on the distance to pollution sources.

[0007] Furthermore, the migration and transformation processes of different water quality parameters in water bodies vary significantly. The spatial arrival order of their peak values ​​can reflect the evolution and migration process of pollutants. However, existing methods usually analyze each parameter in isolation and do not utilize the temporal information contained in the peak misalignment of multiple parameters for comprehensive inversion. This leads to inaccurate judgment of pollution source attributes and location, weak anti-interference ability, and difficulty in meeting the actual needs of high-precision pollution source tracing.

[0008] The information disclosed in the background section is only intended to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0009] The purpose of this application is to provide a method for source region shrinkage and localization of water quality peaks during river navigation, in order to solve the problems mentioned in the background art.

[0010] To achieve the above objectives, this application provides the following technical solution: a method for locating the source region of a peak in river water quality during navigation, comprising the following steps:

[0011] Acquire continuous multi-parameter water quality data and corresponding location information along the river channel. Identify continuous sections with abnormally high water quality parameters as water quality peaks using the continuous multi-parameter water quality data and corresponding location information. Extract the characteristic parameters of the water quality peaks, including peak width and the arrival order of multi-parameter peaks.

[0012] Based on the overall center location of the water quality peak cluster and the corresponding river flow direction data, the range of the initial candidate source area is determined along the counter-current direction.

[0013] Based on the peak width of water quality peaks and the longitudinal dispersion characteristics of the river channel, a mapping relationship between peak width and source area distance is established, the maximum estimated distance of pollution sources is calculated, and areas exceeding the maximum estimated distance are eliminated within the initial candidate source area.

[0014] Based on the arrival order of multiple parameters and the migration and transformation characteristics of each water quality parameter, the migration time of pollutants is determined. Based on the migration time and river flow velocity, areas that do not match the hydrodynamic migration time are screened out. At the same time, interference areas that do not conform to the characteristics of the peak arrival order are excluded, thereby obtaining a further narrowed candidate source area range.

[0015] Based on the gradually narrowed candidate source area, the target area where the pollution source is located is determined, and the spatial range information of the target area is output. Based on the characteristic parameters of the target area and water quality peaks, the pollution source location results are generated and visualized, and relevant data records are output for pollution source tracing analysis.

[0016] Preferably, continuous segments with abnormally high water quality parameters are identified as water quality peak clusters, and characteristic parameters of the water quality peak clusters are extracted. The specific implementation process is as follows:

[0017] Continuous mobile sampling was conducted along the established river course to obtain continuous multi-parameter water quality data and corresponding location information. The sampling points were arranged in chronological order and the water quality parameter values ​​were matched with latitude and longitude information to obtain a data set continuously distributed along the river.

[0018] Based on a continuously distributed dataset along the river, the changes in ammonia nitrogen, permanganate index, turbidity and conductivity along the river were analyzed and anomalies were identified. Adjacent anomalies were defined as anomaly segments and different water quality parameter anomaly segments were superimposed to determine the range of water quality peaks.

[0019] Peak points are extracted from sampling points within the water quality peak cluster, and upper and lower boundary points are determined to obtain the peak width. At the same time, the peak positions of various water quality parameters are compared and sorted to obtain the arrival order of multi-parameter peaks, thereby realizing the extraction of characteristic parameters of water quality peak clusters.

[0020] Preferably, the process of analyzing the changes in ammonia nitrogen, permanganate index, turbidity, and conductivity along the course and identifying anomalies includes: using the median of each water quality parameter over the entire navigation section as a background reference value; and marking the sampling point as an anomaly when the water quality parameter value at the sampling point is higher than the corresponding background reference value by a multiple.

[0021] Preferably, the initial candidate source region range is determined along the counter-current direction, and the specific implementation process is as follows:

[0022] Obtain the data on the overall center position of the water quality peak cluster corresponding to the river section flow direction, convert the flow direction into angular values ​​and divide the direction intervals, count the number of records in each direction interval and determine the dominant flow direction interval;

[0023] Based on the dominant water flow direction interval, the counter-current path is obtained. Starting from the center position, sampling points within the counter-current direction angle range are selected as candidate nodes. The path is extended point by point according to the spatial distance and directional deviation of the nodes.

[0024] Using the path construction results, the sampling points on both sides of the path are connected to obtain a strip-shaped region. Based on the same center position, the downstream region is divided and eliminated, and the upstream strip-shaped region is retained as the initial candidate source area range.

[0025] Preferably, obtaining the countercurrent path based on the dominant water flow direction interval includes: adding 180 degrees to the corresponding angle of the dominant water flow direction interval to obtain the countercurrent direction angle, selecting sampling points located within the countercurrent direction angle range from the set of navigation sampling points as candidate nodes, and selecting nodes to extend the path point by point according to spatial distance and directional deviation.

[0026] Preferably, regions exceeding the maximum estimated distance are removed within the initial candidate source region. The specific implementation process is as follows:

[0027] The peak widths of ammonia nitrogen, permanganate index, turbidity and conductivity within the water quality peak cluster range were extracted and sorted. The peak width with the largest value was selected as the dominant peak width. At the same time, the flow velocity records of the corresponding river section were obtained and the average flow velocity was calculated after removing abnormal data. The maximum estimated distance was determined by matching the upper limit of the source area distance according to the peak width interval.

[0028] Based on the overall center position of the water quality peak cluster, the distance is calculated point by point along the counter-current path from the corresponding sampling point. The distance between adjacent sampling points is accumulated to obtain the cumulative distance value. Sampling points whose cumulative distance does not exceed the maximum estimated distance are marked as retention points. Sampling points that first exceed the maximum estimated distance are taken as boundary points and the area after the boundary point is removed.

[0029] The sampling points corresponding to the retained points are sorted and arranged according to the counter-current path. The distance between adjacent sampling points is judged and continuous segments are divided. The segment with the longest length that includes the sampling point corresponding to the overall center position of the water quality peak is selected as the candidate source area. Sampling points within the range along both sides of the path are connected to complete the distance reduction.

[0030] Preferably, the peak width with the largest value is selected as the dominant peak width, and the corresponding river section flow velocity records are obtained. Records in the flow velocity records that are greater than a certain multiple of the average value of all records or less than a certain multiple of the average value of all records are marked and removed. The average flow velocity value is obtained by arithmetically averaging the remaining flow velocity records, and the maximum estimated distance is determined by matching the upper limit of the source area distance according to the peak width interval.

[0031] Preferably, the candidate source region range is obtained after further shrinkage. The specific implementation process is as follows:

[0032] Extract the peak positions of ammonia nitrogen, permanganate index, turbidity and conductivity within the water quality peak cluster range and project them to the path nodes in the counter-current direction. Determine the arrival order of the multi-parameter peaks according to the arrangement order of the path nodes.

[0033] Based on the correspondence between the arrival order of multi-parameter peaks and the migration and transformation characteristics of each water quality parameter, the migration time is converted and combined with the average flow velocity to determine the migration distance range, and sampling points in the candidate source area whose cumulative path distance is within the migration distance range are screened.

[0034] After filtering, the local peak positions are extracted and the order consistency is judged. The segments with consistent order are retained and the corresponding sampling points are connected as the final candidate source region range, and the further narrowed candidate source region range is obtained.

[0035] Preferably, the arrival order of multiple peak parameters is determined by projecting the peak positions of ammonia nitrogen, permanganate index, turbidity and conductivity within the water quality peak cluster range to the path nodes in the countercurrent direction, and then determining the arrival order according to the arrangement of the path nodes. When multiple peak positions correspond to the same path node or the spacing between path nodes meets the same order condition, the arrival order is determined by the preset priority order of water quality parameters.

[0036] Preferably, based on the gradually narrowed candidate source area, the target area where the pollution source is located is determined. The specific implementation process is as follows:

[0037] The retained sampling points are sorted according to the order of the counter-current path. The first sampling point after sorting is selected as the starting point and adjacent sampling points are connected in sequence to construct a polyline boundary. At the same time, the last sampling point is connected to the starting point to obtain a closed boundary.

[0038] The closed region containing the center of the water quality peak cluster is selected as the unique target region by using the closed boundary. The latitude and longitude coordinates of the boundary points are used to form a polygon coordinate set in sequence. At the same time, the sampling points in the target region are numbered in the order of the counter-current path and the water quality parameter values ​​are bound to the numbers.

[0039] The system uses polygon coordinate sets and numbering sequences to output spatial range information of the target area and generate pollution source location results. At the same time, it visualizes the navigation trajectory, target area boundary, and water quality parameter change curves and outputs relevant data records for pollution source tracing analysis to determine the target area where the pollution source is located.

[0040] The technical effects and advantages provided by this application in the above technical solution are as follows:

[0041] This application constructs a multi-dimensional constraint relationship by introducing water flow direction, peak width, and the arrival order of multiple peak parameters. This unified expression of the spatial distribution characteristics of water quality peaks and the pollutant migration process transforms the original location method, which relied on single concentration changes for inference, into a comprehensive judgment process based on morphological and evolutionary information. This allows for a gradual narrowing of candidate source regions. Through this approach, pollution source location is progressively compressed from a large-scale fuzzy search to specific spatial segments, improving the relevance and stability of the location results and making the source tracing process clearer and more logically supported.

[0042] This application directly extracts and correlates the characteristic parameters of water quality peaks, transforming the migration behavior of pollutants in water bodies into a determinable temporal and spatial constraint relationship. It completes the pollution source area screening process without relying on complex coupled models, effectively reducing the dependence on computing resources. This allows the localization process to complete data processing and result output in a short time, while maintaining consistent judgment logic under complex hydrodynamic conditions, thus meeting the application needs of rapid response and on-site decision-making in sudden pollution events. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0044] Figure 1 This is the overall flowchart of this application;

[0045] Figure 2 A flowchart for identifying continuous segments with abnormally high water quality parameters as water quality peaks in this application and extracting characteristic parameters of the water quality peaks;

[0046] Figure 3 A flowchart illustrating the determination of the initial candidate source region range along the counter-current direction for this application;

[0047] Figure 4 A flowchart for obtaining the further narrowed candidate source region range for this application. Detailed Implementation

[0048] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this application will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0049] like Figures 1 to 4 As shown, this application provides a method for source region contraction and localization of water quality peaks during river navigation, comprising the following steps:

[0050] Step 1: Obtain continuous multi-parameter water quality data and corresponding location information along the river channel. Identify continuous sections with abnormally high water quality parameters as water quality peaks using the continuous multi-parameter water quality data and corresponding location information. Extract the characteristic parameters of the water quality peaks, including peak width and the arrival order of multi-parameter peaks.

[0051] The continuous segments with abnormally high water quality parameters are identified as water quality peak clusters, and the characteristic parameters of these peak clusters are extracted. The specific implementation process is as follows:

[0052] Continuous mobile sampling was conducted along the established river channel, with water quality parameters collected at fixed time intervals, ranging from 5 to 10 seconds. During each sampling, the longitude, latitude, and corresponding time information of the current sampling point were recorded simultaneously, ensuring that each set of water quality parameter data had a unique spatial location identifier. During the sampling process, the mobile speed was controlled to be maintained within the range of 0.5 to 1.5 meters per second, so that the spatial distance between adjacent sampling points was between 2 and 10 meters, thereby ensuring that the data were continuously distributed in the longitudinal direction of the river channel.

[0053] The collected water quality parameters include four categories: ammonia nitrogen, permanganate index, turbidity, and conductivity. During the data processing, all sampling points were arranged in chronological order, and the values ​​of each water quality parameter were matched with the corresponding latitude and longitude information one by one, so that each sampling point contains complete parameter information and spatial information, forming a data set continuously distributed along the river.

[0054] Ammonia nitrogen refers to the nitrogen in water as free ammonia (NH3) and ammonium ions (NH4) + The total amount of nitrogen compounds in the form of nitrogen is an important indicator reflecting the degree of pollution of water bodies by domestic sewage, aquaculture wastewater, or some industrial wastewater. In actual acquisition, continuous measurement is carried out by configuring online ammonia nitrogen sensors while sailing along the river. These sensors are usually based on the ion-selective electrode method or optical colorimetry principle, converting the ammonia nitrogen concentration in the water sample into an electrical or optical signal output. The measurement results are corrected by temperature compensation. The ammonia nitrogen concentration value is recorded in real time at each sampling point and stored synchronously with the corresponding spatial location information, thus forming ammonia nitrogen concentration change data distributed along the river.

[0055] The permanganate index refers to the content of reducing substances in a water sample that can be oxidized by potassium permanganate under certain conditions. It is commonly used to characterize the comprehensive pollution level of organic pollutants and some inorganic reducing substances in water bodies, and is an important indicator reflecting the degree of organic pollution in water bodies. During mobile monitoring, this parameter is obtained by configuring an online permanganate index measurement device. This type of device usually adopts the principle of redox reaction. Under controlled reaction time and conditions, the water sample reacts with potassium permanganate, and the index value is calculated by detecting the amount of potassium permanganate consumed before and after the reaction. The measured data is recorded in a time series format and correlated with the geographical location information of the sampling points, thereby obtaining permanganate index data continuously distributed along the river.

[0056] Turbidity refers to the degree to which suspended particulate matter in water affects the transmission and scattering of light. It reflects the content of sediment, suspended solids, and some organic particles in water and is an important indicator for evaluating the physical pollution status of water bodies. During the acquisition process, turbidity sensors are used to monitor the water body in real time. These sensors typically utilize the principle of light scattering, emitting a light source and receiving the light signal scattered by particulate matter in the water. The scattered light intensity is converted into a turbidity value. Turbidity data is continuously collected at various sampling points during the navigation process, and the collected values ​​are matched with the corresponding spatial coordinates to obtain the turbidity distribution along the river channel.

[0057] Electrical conductivity is a measure of water's ability to conduct electricity, reflecting the total amount of dissolved ions in the water. It is an important indicator for judging changes in the concentration of inorganic salts or some pollutants in water. During mobile monitoring, conductivity sensors are used to continuously measure the water. These sensors typically employ the electrode method, measuring the water's ability to conduct electric current to obtain conductivity values. The measurement results are then compensated and corrected based on water temperature. Conductivity data is acquired at each sampling location and recorded synchronously with the corresponding geographical location and time information, thus forming continuous spatial distribution data of conductivity for subsequent analysis of changes in dissolved substances in the water.

[0058] After obtaining a continuously distributed dataset, the variation of four water quality parameters—ammonia nitrogen, permanganate index, turbidity, and conductivity—was analyzed along the route to identify continuous sections where water quality parameters were abnormally elevated. During the identification process, the median of each water quality parameter in the entire navigation section was used as the background reference value. When the water quality parameter value at a certain sampling point was higher than 1.2 times the corresponding background reference value, the sampling point was marked as an anomaly.

[0059] It's important to note that setting the threshold to 1.2 times is essentially to ensure sensitivity in anomaly detection while avoiding misjudging normal water quality fluctuations as anomalies. During river navigation monitoring, water quality parameters are affected by changes in the natural background, measurement noise, and minor hydrodynamic disturbances. Even without pollution input, parameter values ​​will fluctuate around the background level. If the threshold is set too low, such as close to the background value itself, these normal fluctuations are easily identified as anomalies, leading to an amplified range of abnormal areas. Conversely, if the threshold is set too high, only strong pollution signals will be identified, potentially missing early or weak pollution information. Setting the threshold to 1.2 times the background reference value ensures that values ​​exceed the background level by a certain proportion before being flagged, thus filtering out small fluctuations while retaining meaningful concentration rise characteristics. This makes the identified continuous sections more consistent with the true distribution characteristics of pollution plumes, achieving a balance between stability and sensitivity.

[0060] When three or more consecutive adjacent sampling points are marked as anomalies, and the spatial distance between these sampling points does not exceed 10 meters, these consecutive sampling points are defined as the same anomalous section. For anomalous sections corresponding to different water quality parameters, they are superimposed according to their spatial location. When at least two types of water quality parameters exist in the same spatial range, the spatial range is defined as a water quality peak cluster. The upstream anomalous sampling point in the segment is taken as the starting position, and the downstream anomalous sampling point is taken as the ending position, thus obtaining a water quality peak cluster with a clear spatial boundary.

[0061] After determining the spatial range of the water quality peak cluster, all sampling points within this range are analyzed point by point to extract the characteristic parameters of the water quality peak cluster. During the peak width extraction process, the sampling point with the largest value is selected as the peak point based on the water quality parameter values ​​corresponding to each sampling point within the water quality peak cluster range. 50% of the value corresponding to the peak point is used as a reference threshold. The search proceeds point by point from the peak point upstream. When the water quality parameter value is lower than the reference threshold, the position is determined as the upper boundary point. The search proceeds point by point from the peak point downstream. When the water quality parameter value is lower than the reference threshold, the position is determined as the lower boundary point. The spatial distance between the upper boundary point and the lower boundary point is taken as the peak width of the water quality peak cluster.

[0062] In the process of extracting the peak arrival order of multiple parameters, the values ​​of ammonia nitrogen, permanganate index, turbidity and conductivity within the water quality peak cluster range are compared to determine the sampling point location corresponding to the maximum value of each water quality parameter. These locations are then sorted according to their arrangement in the longitudinal direction of the river channel. When the spatial distance between the peak locations of different water quality parameters is greater than 10 meters, their arrival order is determined from upstream to downstream. When the spatial distance between the peak locations of different water quality parameters is less than 10 meters, these water quality parameters are determined to have arrived simultaneously, thus obtaining a clear peak arrival order of multiple parameters.

[0063] After obtaining the peak width and the arrival order of multiple parameters, the start and end positions, peak widths, and arrival orders of the water quality peak clusters are uniformly organized and correlated with the latitude and longitude information of the original sampling points. This ensures that each water quality peak cluster forms a complete data record. During the data organization process, the spatial range information corresponding to the peak width is associated with the arrangement information corresponding to the arrival order of the multiple parameters. This ensures that the water quality peak cluster contains both spatial range information and sequential information reflecting the characteristics of pollution migration. The above information is continuously stored according to the sampling point order so that the corresponding data can be directly called when conducting source area shrinkage and positioning processes based on water flow direction, peak width, and the arrival order of multiple parameters. This achieves a complete conversion from raw water quality data to characteristic parameters.

[0064] Step 2: Based on the overall center position of the water quality peak cluster and the corresponding river flow direction data, determine the initial candidate source area range along the counter-current direction and eliminate the downstream direction area.

[0065] The initial candidate source region is determined along the counter-current direction. The specific implementation process is as follows:

[0066] Based on the established center position of the water quality peak cluster, the water flow direction data within the corresponding river section is acquired and processed uniformly. In the specific implementation process: the water flow direction is continuously recorded at the center position of the water quality peak cluster, with the flow direction information recorded every 10 seconds for a continuous recording time of no less than 30 minutes. At the same time, the historical flow records of the past 7 days at this position are retrieved, all records are arranged in chronological order, and each water flow direction record is converted into a direction value expressed as an angle. The angle value ranges from 0° to 360°, with due north as 0° and increasing clockwise.

[0067] In this step, the overall center location of the water quality peak cluster is not simply the geometric center or the location of a single peak point, but rather a comprehensive center point that represents the spatial distribution characteristics of the water quality peak cluster. Specifically, it is a representative location determined based on the spatial distribution of all sampling points within the peak cluster. In actual determination, all sampling points between the start and end positions of the peak cluster can be treated as a whole, and values ​​are taken evenly based on the positions of these sampling points in the longitudinal direction of the river channel. This yields the center location located in the middle of the peak cluster's spatial range. This location differs from the peak center that only reflects the maximum concentration and also from the geometric center calculated solely based on the boundary. Instead, it is a reference point that takes into account both the spatial extension range of the peak cluster and the continuity of data distribution. It is used to uniformly characterize the overall position of the water quality peak cluster in the river channel and serves as a spatial benchmark for subsequent flow direction determination and source area back-calculation.

[0068] In the direction classification process, 360° is divided into 8 direction intervals, each with a range of 45°, corresponding to north, northeast, east, southeast, south, southwest, west, and northwest respectively. Each record is assigned to the corresponding interval, and the number of records in each interval is counted. The interval with the most records is determined as the dominant flow direction interval. When two intervals have the same number of records, the interval with the smallest difference in angle from the current real-time recorded direction is selected as the dominant direction. This dominant direction interval is then converted into the corresponding river flow direction description, so that the overall center position of the water quality peak cluster has a uniquely determined flow direction attribute.

[0069] Subsequently, starting from the overall center of the water quality peak cluster, the countercurrent path is obtained based on the determined dominant flow direction interval. In the specific implementation process: the angle corresponding to the dominant flow direction interval is added by 180° to obtain the countercurrent angle, and this angle is used as the target direction for path extension. Starting from the center, all sampling points within ±22.5° of the countercurrent angle are selected as candidate nodes from the set of mobile sampling points. Among the candidate nodes, the sampling point with the smallest spatial distance from the current node is selected as the next path node. When there are multiple candidate nodes with the same distance, the deviation values ​​between these candidate nodes and the countercurrent angle are further compared, and the sampling point with the smallest deviation value is selected as the next node. The path is extended upstream point by point according to this rule, so that the path always advances continuously along the countercurrent direction.

[0070] During the path extension process, when the distance between adjacent nodes exceeds 10 meters, the current direction of extension is stopped, and candidate nodes that meet the angle range are re-selected around the current node to continue the extension, so that the path remains continuous in space. When the total length of the path reaches 1500 meters, the extension stops, and the sequence of all nodes from the center position to the end of the path is recorded.

[0071] After the path is constructed, sampling points within 20 meters of the path centerline are selected on both sides of the path, using the path as the centerline. The spatial regions corresponding to these sampling points are connected to form a strip region, which is used as the initial candidate source region range.

[0072] After delineating the initial candidate source area in the countercurrent direction, the center of the same water quality peak cluster is used as the dividing point to clearly define and eliminate the downstream area. In the specific implementation process: the angle of the dominant water flow direction is taken as the downstream direction. Starting from the center, sampling points within ±22.5° of the downstream direction angle are selected from the sampling point set. Following the same node selection rules as the countercurrent path construction, the path is extended downstream point by point. Extension stops when the total path length reaches 1500 meters, and all nodes corresponding to this path are recorded. Based on this path... Sampling points were selected within 20 meters of the center line on both sides of the path. The areas corresponding to these sampling points were connected to form a downstream strip. This downstream strip was then removed from the river channel space, and only the strip corresponding to the upstream direction was retained as a candidate source area. This ensured that the candidate source area was located only upstream of the center of the water quality peak cluster. This unidirectional constraint on the water flow direction achieved the initial shrinkage of the pollution source area, providing a uniquely determined spatial range for further shrinkage based on peak width and the arrival order of multiple peak parameters.

[0073] Step 3: Based on the peak width of the water quality peak cluster and the longitudinal dispersion characteristics of the river channel, establish the mapping relationship between the peak width and the source area distance, calculate the maximum estimated distance of the pollution source, and remove areas exceeding the maximum estimated distance within the initial candidate source area.

[0074] Regions exceeding the maximum estimated distance are removed from the initial candidate source region. The specific implementation process is as follows:

[0075] Based on the obtained water quality peak range and corresponding peak width, the peak widths are uniformly organized and their correspondence with the longitudinal dispersion characteristics of the river channel is established. Specifically, the peak width values ​​of four water quality parameters—ammonia nitrogen, permanganate index, turbidity, and conductivity—are extracted one by one from the water quality peak range. The peak widths corresponding to these four parameters are then sorted according to their numerical values, and the peak width ranking first in the sorting results is selected as the dominant peak width of the current water quality peak. Simultaneously, flow velocity records for the past twelve months in the river segment containing the water quality peak are obtained. All flow velocity records are arranged chronologically, and outlier data is removed according to fixed rules.

[0076] When a flow velocity record value is greater than 1.5 times the average of all records or less than 0.5 times the average of all records, the record is marked as an anomaly and removed from the dataset. After the anomaly record removal is completed, the arithmetic mean of the remaining flow velocity records is calculated to obtain the average flow velocity value of the river section.

[0077] Based on this, the peak width is classified according to the fixed interval division rule, dividing the peak width into continuous and non-overlapping intervals, specifically three intervals: greater than or equal to 10 meters and less than 30 meters, greater than or equal to 30 meters and less than 60 meters, and greater than or equal to 60 meters and less than 120 meters. Each interval corresponds to a uniquely determined upper limit of the source area distance, where greater than or equal to 10 meters and less than 30 meters corresponds to 300 meters, greater than or equal to 30 meters and less than 60 meters corresponds to 600 meters, and greater than or equal to 60 meters and less than 120 meters corresponds to 1200 meters. When the dominant peak width falls into a certain interval, the corresponding upper limit of the distance is selected as the maximum estimated distance of the pollution source according to the unique attribution rule of the above interval boundary, thereby establishing a non-overlapping and unambiguous mapping relationship between the peak width and the source area distance.

[0078] It should be noted that the longitudinal dispersion characteristic of a river channel refers to the physical process by which pollutants gradually diffuse and stretch in the longitudinal direction during their propagation along the river's flow direction due to factors such as differences in water velocity, turbulent diffusion, and uneven velocity distribution across the cross-section. Specifically, when pollutants enter the water body, while they are transported as a whole, differences in flow velocity at different water layers and locations cause the pollutant plume to accelerate at its leading edge and lag at its trailing edge during its movement. This results in the gradual widening of the initially concentrated pollutant distribution, manifested as a decrease in peak concentration and an expansion of its spatial range. This degree of expansion along the river channel is the longitudinal dispersion characteristic. This characteristic reflects the diffusion intensity and mixing degree of pollutants during migration and is usually related to river velocity, water depth changes, cross-sectional morphology, and the turbulent state of the water body. It is an important basis for establishing the mapping relationship between peak width and source region distance.

[0079] After obtaining the maximum estimated distance to the pollution source, the overall center position of the water quality peak cluster is used as the spatial starting point. This maximum estimated distance is applied to the initial candidate source area range determined in step two. In the specific implementation process: the counter-current path constructed in step two is used as the distance calculation path, and the arrangement order of the sampling points in this path is used as the distance calculation order. Starting from the sampling point corresponding to the overall center position of the water quality peak cluster, the distance is calculated point by point along the path direction. During the distance calculation process, the spatial distance between two adjacent sampling points is converted into a planar distance according to the difference in latitude and longitude, and this distance is accumulated segment by segment, so that each Each sampling point corresponds to a cumulative distance value. When the cumulative distance is less than or equal to the maximum estimated distance, the corresponding sampling point is marked as a retention point. When the cumulative distance exceeds the maximum estimated distance for the first time, the sampling point is marked as a boundary point, and path extension stops. All sampling points after the boundary point are marked as removal points. At the same time, the entire area between the boundary point and the end of the path is defined as an out-of-range area and removed from the initial candidate source area range, with only the path area from the overall center of the water quality peak cluster to the boundary point retained, so that the candidate source area is spatially limited to the maximum estimated distance range.

[0080] After completing the region selection based on the maximum estimated distance, the retained regions undergo continuous organization and spatial expansion processing. Specifically:

[0081] All sampling points marked as reserved points are arranged sequentially along the counter-current path. The spatial distance between adjacent sampling points is judged point by point. When the distance between adjacent sampling points is no more than 10 meters, these sampling points are grouped into the same continuous segment. When the distance between adjacent sampling points is greater than 10 meters, the location is used as the segment boundary point, thereby dividing the reserved points into multiple continuous segments. After the division is completed, the segment containing the sampling point corresponding to the overall center position of the water quality peak is selected as the only valid candidate source area. When multiple segments contain the center position, the segment containing the center position and with the longest length is selected as the final segment, and all other segments are removed.

[0082] After determining the final segment, the path of the segment is used as the center line. Sampling points within a distance of no more than 20 meters are selected on both sides of the center line. The spatial regions corresponding to these sampling points are connected according to their adjacency relationship to form a continuous strip-shaped region. This yields the final distance shrinkage candidate source region range, which is constrained by the maximum estimated distance in the longitudinal direction of the river channel and forms a continuous coverage area in the lateral direction. This completes the spatial shrinkage process based on the mapping relationship between peak width and source region distance.

[0083] Step 4: Based on the arrival order of multiple parameters and the migration and transformation characteristics of each water quality parameter, determine the pollutant migration time, and screen out areas that do not match the hydrodynamic migration time according to the migration time and river flow velocity. At the same time, exclude interference areas where the peak arrival order does not conform to the characteristics, and then obtain the further narrowed candidate source area range.

[0084] The specific implementation process for obtaining the further narrowed candidate source region range is as follows:

[0085] The peak positions of four water quality parameters—ammonia nitrogen, permanganate index, turbidity, and conductivity—within the water quality peak cluster are uniformly extracted. These peak positions are then spatially sorted according to the countercurrent path determined in step three. During the sorting process, the peak positions of each water quality parameter are projected onto the nearest path node on the countercurrent path, and the order of the path nodes in the path is used as the sorting basis. When two peak positions correspond to the same path node, they are considered to be the same position. When the distance between the path nodes corresponding to two peak positions is no more than 10 meters, they are determined to arrive in the same order. When the distance between the path nodes is greater than 10 meters, the arrival order is determined according to the path nodes from the position closer to the center of the water quality peak cluster to the position farther away from the center. When multiple water quality parameter peak positions are equidistant from the path nodes, they are sorted according to the preset priority order of the water quality parameters. The priority order is set as conductivity first, permanganate index first, turbidity first, and turbidity first, ammonia nitrogen first, thus obtaining a uniquely determined multi-parameter peak arrival order.

[0086] After obtaining a unique and definite peak arrival order for multiple parameters, a fixed correspondence is established between this order and the migration and transformation characteristics of each water quality parameter. In specific implementation: conductivity is set as a reference parameter that does not decay, permanganate index is set as a slowly changing parameter, turbidity is set as a sedimentation changing parameter, ammonia nitrogen is set as a rapidly changing parameter, and the peak arrival order for multiple parameters is converted into a migration time level.

[0087] When the peak positions of the four water quality parameters are determined to arrive in the same order, the migration time is set to 10 minutes. When conductivity and permanganate index arrive in the same order and turbidity and ammonia nitrogen are downstream, the migration time is set to 30 minutes. When conductivity is upstream and ammonia nitrogen is downstream and the distance between each peak position is greater than 10 meters, the migration time is set to 60 minutes. When the maximum distance between peak positions exceeds 150 meters, the migration time is set to 120 minutes. The arrival order of the multi-parameter peaks is converted into a unique corresponding migration time value through the above fixed rules.

[0088] After obtaining the migration time value, the migration time is combined with the average flow velocity determined in step three to determine the corresponding migration distance range. In the specific implementation process: the migration time is multiplied by the average flow velocity to obtain the migration distance, and this migration distance is used as the center value. A distance interval is formed by extending the range above and below it by 20%. This distance interval is applied to the candidate source area range retained in step three. During the screening process, the cumulative distance of the path is calculated point by point along the counter-current path determined in step two, starting from the overall center position of the water quality peak. When the cumulative distance corresponding to a certain sampling point is less than the lower limit of the distance interval or greater than the upper limit of the distance interval, the sampling point is marked as a mismatch point and removed from the candidate source area. Only sampling points whose cumulative path distance is within the distance interval are retained, thereby completing the regional screening based on the matching of migration time and river flow velocity.

[0089] After distance matching screening, peak arrival order consistency screening is carried out in the remaining area. In the specific implementation process: the local peak positions of ammonia nitrogen, permanganate index, turbidity and conductivity are extracted for each continuous segment in the remaining area, and the local peak arrival order is determined according to the same sorting rules as mentioned above. When the local order is consistent with the determined overall peak arrival order, the segment is marked as a valid segment. When the position order of any water quality parameter in the local order is inconsistent with the overall order, the segment is marked as an interfering segment and removed.

[0090] After the screening is completed, all sampling points marked as valid segments are connected continuously in the order of the path, and the connected area is used as the final candidate source area range, thereby realizing the source area attribute fine shrinkage based on the multi-parameter peak arrival order and migration time constraints.

[0091] After completing the peak arrival sequence consistency screening and removing interfering segments, all sampling points marked as valid segments are continuously connected according to the order of the counter-current path. The connected spatial area is then integrated as a whole, and the parts where the distance between adjacent sampling points does not exceed the preset interval are merged into the same continuous area, thus obtaining a spatially continuous and unique area. This area is then used as the candidate source area range after being screened step by step by water flow direction constraints, peak width constraints, and multi-parameter peak arrival sequence constraints. This is the gradually narrowed candidate source area range, which is used for the final determination of the target area of ​​the subsequent pollution source.

[0092] Step 5: Based on the gradually narrowed candidate source area, determine the target area where the pollution source is located and output the spatial range information of the target area; generate pollution source location results information and visualize them for the characteristic parameters of the target area and water quality peaks, and output relevant data records for pollution source tracing analysis.

[0093] Based on the gradually narrowed candidate source area, the target area where the pollution source is located is determined. The specific implementation process is as follows:

[0094] After completing the stepwise contraction based on the direction of water flow, peak width, and the arrival order of multiple peak parameters, the spatial boundary is constructed for all the sampling points retained within the candidate source area. In the specific implementation process: all the retained sampling points are sorted according to their arrangement order on the counter-current path in step two, and the first sampling point after sorting is used as the starting point. Adjacent sampling points are connected in sequence to form a broken line boundary. When the last sampling point is connected, the last sampling point is connected to the starting point to form a closed boundary. When the distance between adjacent sampling points is greater than 10 meters, supplementary points are inserted between the two points according to the rule of not more than 10 meters apart and participate in the connection to ensure the continuity of the boundary.

[0095] After the closed boundary is formed, the closed area is used as the single target area in the candidate area set. When there are multiple closed areas, the closed area containing the overall center of the water quality peak is selected as the unique target area, and all other areas are eliminated, thus obtaining the target area where the pollution source is located. The latitude and longitude coordinates of all boundary points of the area are used to form a polygon coordinate set in sequence, which serves as the spatial range information of the target area.

[0096] After obtaining the spatial range information of the target area, the characteristic parameters of the target area and the water quality peak cluster are matched one by one. The specific process is as follows: all sampling points in the target area are numbered according to their order in the counter-current path, starting from 1 and increasing sequentially, so that each sampling point has a unique number identifier. At the same time, the values ​​of ammonia nitrogen, permanganate index, turbidity and conductivity corresponding to the sampling point are bound to the number. The peak width of the water quality peak cluster, the arrival order of the peaks of multiple parameters and the water flow direction information are uniformly associated with the number sequence, so that each number corresponds to a complete set of data records, thereby forming a unified data set containing spatial location, characteristic parameters and water quality data.

[0097] After completing the data association and organization, the pollution source location results are output in a standardized manner. Specifically, the polygon coordinate set of the target area is output as a coordinate list in numerical order, and the water quality parameter values ​​corresponding to the number sequence are output as a data list in the same order. At the same time, the peak width of the water quality peak cluster, the arrival order of the multi-parameter peaks, and the water flow direction information are appended to the coordinate list in a fixed text format. The output order is: water flow direction information, peak width information, multi-parameter peak arrival order information, coordinate list information, and sampling point data list information, thus forming pollution source location results with a unified structure and fixed order.

[0098] After outputting the pollution source location results, the results are visualized and the data is recorded. In the specific implementation process: the navigation trajectory is connected sequentially according to the sampling point number to form a continuous trajectory line, and the sampling point position is marked on the trajectory line with the number as the index. The polygon coordinates of the target area are drawn as a closed boundary according to the same number order and covered on the trajectory line. At the same time, curves of ammonia nitrogen, permanganate index, turbidity and conductivity as a function of the path are drawn at the corresponding position of the trajectory line, and the peak position of each parameter and its arrival sequence number are marked on the curve. All graphic elements are displayed in sequence according to the trajectory line, the target area boundary and the parameter curve.

[0099] In terms of data recording, the original sampling data, peak extraction data, feature parameter data, and sampling point numbering information retained and removed during each shrinkage process are organized in chronological order and output in a unified data file format, thereby achieving complete recording of all data required for pollution source tracing analysis.

[0100] The foregoing has only described certain exemplary embodiments of this application by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of this application. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of this application.

Claims

1. A method for locating the source region of a peak water quality cluster during river navigation, characterized in that, Includes the following steps: Acquire continuous multi-parameter water quality data and corresponding location information along the river channel. Identify continuous sections with abnormally high water quality parameters as water quality peaks using the continuous multi-parameter water quality data and corresponding location information. Extract the characteristic parameters of the water quality peaks, including peak width and the arrival order of multi-parameter peaks. Based on the overall center location of the water quality peak cluster and the corresponding river flow direction data, the initial candidate source area range is determined along the counter-current direction. Based on the peak width of water quality peaks and the longitudinal dispersion characteristics of the river channel, a mapping relationship between peak width and source area distance is established, the maximum estimated distance of pollution sources is calculated, and areas exceeding the maximum estimated distance are eliminated within the initial candidate source area. Based on the arrival order of multiple parameters and the migration and transformation characteristics of each water quality parameter, the migration time of pollutants is determined. Based on the migration time and river flow velocity, areas that do not match the hydrodynamic migration time are screened out. At the same time, interference areas that do not conform to the characteristics of the peak arrival order are excluded, thereby obtaining a further narrowed candidate source area range. Based on the gradually narrowed candidate source area, the target area where the pollution source is located is determined, and the spatial range information of the target area is output. Based on the characteristic parameters of the target area and water quality peaks, the pollution source location result information is generated.

2. The method for locating the source region contraction of a river channel water quality peak as described in claim 1, characterized in that, The continuous range of abnormally elevated water quality parameters is identified as water quality peak clusters, and the characteristic parameters of these peak clusters are extracted. The specific implementation process is as follows: Continuous mobile sampling was conducted along the established river course to obtain continuous multi-parameter water quality data and corresponding location information. The sampling points were arranged in chronological order and the water quality parameter values ​​were matched with latitude and longitude information to obtain a data set continuously distributed along the river. Based on a continuously distributed dataset along the river, the changes in ammonia nitrogen, permanganate index, turbidity and conductivity along the river were analyzed and anomalies were identified. Adjacent anomalies were defined as anomaly segments and different water quality parameter anomaly segments were superimposed to determine the range of water quality peaks. Peak points are extracted from sampling points within the water quality peak cluster range, and upper and lower boundary points are determined to obtain the peak width. At the same time, the peak positions of various water quality parameters are compared and sorted to obtain the arrival order of multi-parameter peaks, thereby realizing the extraction of characteristic parameters of water quality peak clusters.

3. The method for locating the source region contraction of a river channel water quality peak as described in claim 2, characterized in that, The process of analyzing the changes in ammonia nitrogen, permanganate index, turbidity, and conductivity along the route and identifying anomalies includes: using the median of each water quality parameter over the entire navigation section as the background reference value; and marking the sampling point as an anomaly when the water quality parameter value at the sampling point is higher than the corresponding background reference value by a multiple.

4. The method for locating the source region contraction of a river channel water quality peak as described in claim 2, characterized in that, The initial candidate source region is determined along the counter-current direction. The specific implementation process is as follows: Obtain the data on the overall center position of the water quality peak cluster corresponding to the river section flow direction, convert the flow direction into angular values ​​and divide the direction intervals, count the number of records in each direction interval and determine the dominant flow direction interval; Based on the dominant water flow direction interval, the counter-current path is obtained. Starting from the center position, sampling points within the counter-current direction angle range are selected as candidate nodes. The path is extended point by point according to the spatial distance and directional deviation of the nodes. Using the path construction results, the sampling points on both sides of the path are connected to obtain a strip-shaped region. Based on the same center position, the downstream region is divided and eliminated, and the upstream strip-shaped region is retained as the initial candidate source area range.

5. The method for locating the source region contraction of a river channel water quality peak as described in claim 4, characterized in that, Obtaining the countercurrent path based on the dominant water flow direction interval includes: adding 180 degrees to the corresponding angle of the dominant water flow direction interval to obtain the countercurrent direction angle; selecting sampling points within the countercurrent direction angle range from the set of navigation sampling points as candidate nodes; and selecting nodes to extend the path point by point according to spatial distance and directional deviation.

6. The method for locating the source region of a river navigation water quality peak cluster according to claim 4, characterized in that, Regions exceeding the maximum estimated distance are removed from the initial candidate source region. The specific implementation process is as follows: The peak widths of ammonia nitrogen, permanganate index, turbidity and conductivity within the water quality peak cluster range were extracted and sorted. The peak width with the largest value was selected as the dominant peak width. At the same time, the flow velocity records of the corresponding river section were obtained and the average flow velocity was calculated after removing abnormal data. The maximum estimated distance was determined by matching the upper limit of the source area distance according to the peak width interval. Based on the overall center position of the water quality peak cluster, the distance is calculated point by point along the counter-current path from the corresponding sampling point. The distance between adjacent sampling points is accumulated to obtain the cumulative distance value. Sampling points whose cumulative distance does not exceed the maximum estimated distance are marked as retained points. Sampling points that first exceed the maximum estimated distance are taken as boundary points and the area after the boundary point is removed.

7. The method for locating the source region contraction of a river channel water quality peak as described in claim 6, characterized in that, The peak width with the largest value is selected as the dominant peak width, and the corresponding river section flow velocity records are obtained. Records with flow velocity values ​​greater than a certain multiple of the average value of all records or less than a certain multiple of the average value of all records are marked and removed. The average flow velocity value is obtained by arithmetically averaging the remaining flow velocity records. The maximum estimated distance is determined by matching the upper limit of the source area distance according to the peak width interval.

8. The method for locating the source region contraction of a river channel water quality peak as described in claim 6, characterized in that, The specific implementation process for obtaining the further narrowed candidate source region range is as follows: Extract the peak positions of ammonia nitrogen, permanganate index, turbidity and conductivity within the water quality peak cluster range and project them to the path nodes in the counter-current direction. Determine the arrival order of the multi-parameter peaks according to the arrangement order of the path nodes. Based on the correspondence between the arrival order of multi-parameter peaks and the migration and transformation characteristics of each water quality parameter, the migration time is converted and combined with the average flow velocity to determine the migration distance range, and sampling points in the candidate source area whose cumulative path distance is within the migration distance range are screened. After filtering, the local peak positions are extracted and the order consistency is judged. The segments with consistent order are retained and the corresponding sampling points are connected as the final candidate source region range, and the further narrowed candidate source region range is obtained.

9. The method for locating the source region contraction of a river channel water quality peak as described in claim 8, characterized in that, The arrival order of multiple parameters is determined by projecting the peak positions of ammonia nitrogen, permanganate index, turbidity and conductivity within the water quality peak cluster to the path nodes in the counter-current direction, and then determining the order of arrival according to the arrangement of the path nodes. When multiple peak positions correspond to the same path node or the spacing between path nodes meets the same order condition, the arrival order is determined by the preset priority order of water quality parameters.

10. The method for locating the source region contraction of a river channel water quality peak according to claim 8, characterized in that, Based on the gradually narrowed candidate source area, the target area where the pollution source is located is determined. The specific implementation process is as follows: The retained sampling points are sorted according to the order of the counter-current path. The first sampling point after sorting is selected as the starting point and adjacent sampling points are connected in sequence to construct a polyline boundary. At the same time, the last sampling point is connected to the starting point to obtain a closed boundary. The closed region containing the center of the water quality peak cluster is selected as the unique target region by using the closed boundary. The latitude and longitude coordinates of the boundary points are used to form a polygon coordinate set in sequence. At the same time, the sampling points in the target region are numbered in the order of the counter-current path and the water quality parameter values ​​are bound to the numbers. The system uses polygon coordinate sets and numbering sequences to output spatial range information of the target area and generate pollution source location results. At the same time, it visualizes the navigation trajectory, target area boundary, and water quality parameter change curves and outputs relevant data records for pollution source tracing analysis to determine the target area where the pollution source is located.

Images